MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

So I've been trying to understand a proof of the fact that $H^p(M, \mathcal F)=0$ whenever $p\geq 1$ from page 42 of the book "Principles of Algebraic Geometry" by Griffiths and Harris.

The proof is carried out for the sheaf ${\mathcal a}^{r,s}$ of $C^{\infty}$ forms of type $(r,s)$ on $U$ and there is a remark after the proof of how one can prove the result for an arbitrary fine sheaf.

According to them, a fine sheaf is a sheaf for which we have maps $\eta_\alpha:\mathcal F (U_{\alpha})\rightarrow \mathcal F (U)$ for any $U=\cup U_{\alpha}$, such that the support of $(\eta_{\alpha}\sigma)$ is contained in $U_\alpha$ and $\sum {\eta_{\alpha}} (\sigma| _{U_\alpha})=\sigma$ for $\sigma\in \mathcal F (U)$.

So is $(\eta_{\alpha}\sigma)$ a map from the manifold $M$ to $\mathbb R$? I don't understand how $(\eta_{\alpha}\sigma)$ has $M$ as its domain.

Secondly how can they write $(\eta_{\alpha}\sigma)$ when $\sigma\in \mathcal F (U)$? Isn't the domain of $\eta_\alpha$ the group ${\mathcal F }(U_\alpha)$?

Thirdly how does this formulation apply to the case when $\mathcal F$ is the sheaf ${\mathcal a}^{(r,s)}$? We would need maps $\eta_{\alpha}: {\mathcal a}^{(r,s)}(U_{\alpha})\rightarrow {\mathcal a}^{(r,s)}(U)$. But indstead the proof just uses the partion of unity $C^{\infty} $ maps $\rho_{\alpha}$ subordinate to some open cover $\{U_\alpha\}$. So it seems that they haven't used the fact that ${\mathcal a}^{r,s}$ is a fine sheaf.

share|cite|improve this question
miss a "$" in the third paragraph? It doesn't display correctly here...And I think in order to imitate the partition of unity, G&H abuses the language here. Probably you need a reference for a more detailed version about these. I learned these stuff from Voisin's hodge theory and complex algebraic geometry. – Honglu May 27 '11 at 6:27
I've tried to fix up the latex - please correct if I've destroyed the meaning of what was intended. – David Roberts May 27 '11 at 7:20
Yes, thank you David. That's what I intended. – Nakhoul May 27 '11 at 7:27
Yes the problem with learning from Grifiths and Harris is that they constantly abuse notation, making a lot of what they say somewhat elusive. – Nakhoul May 27 '11 at 7:33
up vote 2 down vote accepted

About your third question. Using partition of unity $(\rho_\alpha)_\alpha$ subordinate to $(U_\alpha)_\alpha$ for $\mathcal F$ you can define maps $\eta_\alpha:\mathcal F(U_\alpha)\to\mathcal F(U)$ by taking $s\in\mathcal F(U_\alpha)$ to $\rho_\alpha s$.

About the first two points it seems to me that $(\eta_\alpha\sigma):=\eta_\alpha(\sigma_{|U_\alpha})$.

EDIT: to summarize my answer a fine sheaf is, "by definition", a sheaf having partition of unity.

share|cite|improve this answer
Thanks. Though I still don't understand what it means to say that $support(\eta_\alpha \sigma)\subset U_\alpha$. I can understand what it means in the case where $\mathcal F$ is the sheaf of $C^\infty$ functions, but when it comes to an arbitrary element $\sigma$ of some section $\mathcal F (U)$, I cannot fathom what it means for $\sigma$ to only be non zero on $U_\alpha$, where $U_\alpha$ is some element of a open cover of $M$. – Nakhoul May 27 '11 at 13:28
For a section $\tau\in \mathcal F(U)$ (e.g. $\tau=(\eta_\alpha\sigma)$), the complement $C$ of its support is defined as follows: a point $x\in U$ lies in $C$ iff there exists an open neighborhood $V$ of $x$ in $U$ such that $tau_{|V}=0$. – DamienC May 27 '11 at 14:56
Thanks so much for clearing that up. – Nakhoul May 27 '11 at 15:34

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.